460 Journal of Medicinal Chemistrj, 1978, Vol. 21, No. 5
Rankin, Riley, Murphj
3-Aryl- and 3-Hydroxy-3-aryloctahydropyrido[2,1-c][1,4]oxazines. Synthesis, Stereochemistry, and Central Nervous System Pharmacological Actions Gary 0. Rankin,' Thomas N. Riley," and James C. Murphy2 Department of Medicinal Chemistry, School of Pharmacy, University of Mississippi, University, Mississippi 38677. Received January 11, 1978 A series of substituted 3-aryl- and 3-hydroxy-3-aryloctahydropyrido[2,1-c] [ 1,4]oxazines has been synthesized for purposes of investigating potentially useful CNS pharmacological actions of this novel heterocyclic system. The preferred conformation of the bicyclic system of the parent compounds, 1 and 2, has been shown to be trans with equatorial orientation of the 3-phenyl substituent in each compound. The various substituted aryl analogues of this series are depressants of motor activity in mice, and certain analogues exhibit significant anticonvulsant and appetite suppressant activity in test animals.
We report here the results of a structure-activity relationship study of the influence of phenyl substituents on the central nervous system (CNS) actions of 3-phenyl(1) and 3-hydroxy-3-phenyloctahydropyrido[2,1-c] [ 1,4]oxazines (2). A previous report" describing the CNS 9
6
Scheme I
OH
2, 22-34
1
5
4
1
X
1,X=H 2 , X = OH
3
4, R = H (phenmetrazine) 5, R = CH, (phendimetrazine)
6
actions of 1 and 2 indicated that these compounds possess a depressant action demonstrated by a reduction of locomotor activity of mice. The potential for therapeutic utility of the central actions of the octahydro[2,1-c][1,4]oxazines is indicated by the various CNS pharmacological actions reported for related 2-arylmorpholines, such as psychomotor stimulation and anorexia (3-5)4,5 and anticonvulsant activity (6).6 Synthesis. The desired phenyl-substituted analogues of 1 and 2 were prepared according to the pathway illustrated in Scheme I. Treatment of 2-piperidinemethanol with the appropriately substituted 2-bromoacetophenone provided quantitative yields of the hemiketal product (analogues of 2). There was no evidence to suggest the presence of the keto alcohol tautomer. Reductive cleavage of the hemiketal provided diol intermediates which were cyclized to the appropriate analogues of 1 under acidic conditions. Optimal conditions for conversion of the hemiketals to the diols involved the use of 10% Pd(OH), on C [10:1 (w/w) hemiketal to catalyst], 1 N H2S04as solvent, and 48 psi of H2 for 24 h. The haloaryldiol intermediates were prepared from the appropriate hemiketal by NaBH, reduction to avoid aromatic dehalogenation which occurs under catalytic reductive conditions. Cyclization of diol intermediates was accomplished in refluxing 48% HBr for a 2-h reaction period. The methoxyaryl analogues 14-17 were synthesized from the appropriate diols by refluxing in 50% H2S04(20-45 min) so as to avoid 0-demethylation. Other cyclization procedures, including refluxing in AczO or heating in either Me2S0 or 0022-2623/78/ 1821-0460$01.00/0
1 , 7-21
P0Cl3, were evaluated and found to be unsatisfactory as compared to the HBr or H2S04catalyzed procedures (see Table I). Stereochemical Studies. The bicyclic octahydropyrido[2,1-c][ 1,4]oxazine system provides a more rigid stereochemical framework than found in the 2-arylmorpholines. Hence, an analysis of the stereochemistry of 1 and 2 may provide useful information regarding the steric aspects of the CNS actions of these compounds and their substituted phenyl analogues. The octahydropyrido[2,1-c][ 1,410xazine system can display interconversion of cis-trans ring fused chair conformers via inversion about the bridgehead N. However, the presence of strong Bohlmann bands (2700-2800 cm-') in the IR spectra of 1 and 2 suggests a preference for the trans-fused conformers. The presence of Bohlmann bands in the IR spectra of related bicyclic systems serves as a reliable indicator for trans stere~chemistry.~Reports of the preferred conformations of perhydro[ 1,2-c][ 1,3]oxazine8 and 3-substituted quinoli~idines~ have indicated a trans fusion for these molecules. One would expect contributions of trans fusion to the preferred conformations of 1 and 2 to be even greater than for these related systems because of the absence of nonbonded interactions between C(lO)-Hax and C(2) substituents in the octahydropyrido[ 2,1-c][ 1,410xazines. Evidence for a preferred equatorial orientation of the phenyl substituent in 1 was provided by the appearance of an ABX pattern for the C(4)-methylene protons and the C(3)-H, in the NMR spectrum, 6 2.1 and 3.1 and 6 4.65 ( J = 3 and 11 Hz), respectively. A singlet (6 7.25) for the aromatic protons of 1 is also indicative of equatorial P h in that it is consistent with the results obtained in comparing the NMR spectra of 3(e)- and 3(a)-phenylq~inolizidine.~ Analysis of pseudocontact deshielded NMR spectra of 1 utilizing the lanthanide shift reagent Eu(fod)310supports the configurational assignment for this compound. The signals for the CY protons of the oxazine ring, HI,,!, H,(,,, and H4(+ experienced significantly greater deshielding than the p protons as the ratio of Eu(fod)3to 1 is increased. These data are interpreted on the basis of the reasonable assumption that the lanthanide shift reagent is preferentially complexed with the basic N 1978 American Chemical Society
Journal of Medicinal Chemistry, 1978, Vol. 21, No. 5 461
Substituted Octahydropyrido[2,1 -c][l,4]oxatines
Table I. 3-Aryl- and 3-Hydroxy-3-aryloctahydropyrido[ 2,l-c] [ 1,410xazines
upR X
Compd
R
X
MP, "C 212-214 206-208 255-257 275-277 222-224 218-219 239-241 220-222 173-174 191-193 209-210 257-259 199-200 243-244 181-183 264-265 166-167 155-157 168-170 177-179 163-164 172-174 198-199 160-161 140-141 1 51-153 161-163 154-155 153-1 5 4 202-204
H 4-F 8 4-C1 9 4-Br 2-CF3 10 3-CF3 11 4-CF3 12 4-CH3 13 3-OCH3 14 15 4-OCH, 3&-(OCH3 16 3,4,5-(OCH, 13 17 18 3-OH 4-OH 19 3-OAc 20 4-NHAc 21 H 2 4-F 22 23 4421 4-Br 24 25 2-CF3 26 3-CF3 27 4-CF3 4-CH3 28 29 3-OCH3 4-OCH, 30 3,4-(OCH3 31 3,4,5-(OCH,), 32 3-OAc 33 34 4-NO, Analyzed for C, H, and N. a Ace = acetone, PhH = benzene. 1 7
H H H H H H H H H H H H H H H H OH OH OH OH OH OH OH OH OH OH OH OH OH OH
atom of 1.'l Evidence for a 3(a)-hydroxy-3(e)-phenyl configuration for 2 was obtained from an examination of
1
the IR and NMR spectra of this compound. A significant amount of intramolecular H-bonded species was noted in the IR spectrum (dilution studies) of 2. Only the trans conformer of 2 having C(3)-OHa, is capable of forming intramolecular H bonds. The appearance of the signal for H1(,) downfield relative to that for H1(,) in the NMR spectrum of 2 can be explained in terms of deshielding of the former proton by the C(3)-OH,. Pseudocontact deshielded NMR spectra of 2 utilizing Eu(fod)a indicated greater deshielding of the a protons of the oxazine ring of this compound, providing further evidence for the axial orientation of the C(3)-OH since preferential interaction of the shift reagent and 2 would be expected to occur a t the OH function.'l The possibility that a tautomeric equilibrium could contribute significantly to the structure of the hemiketal 2 was considered. Tautomeric interOH
T Ph
L
conversion could conceivably give rise to configurational inversion a t C(3). Examination of the solid (KBr) and solution (CC14)IR spectra of 2 failed to reveal any carbonyl
Recrystn solventa Ace Ace-Et,O Ace Ace PhH-Et, 0 PhH Ace-Et,O Ace-Et,O Ph H PhH PhH-Et,O Ace Et OH-Et, 0 EtOH-Et,O Ace-Et,O Ace Ace-Et,O Ace-Et,O Ace Ace EtOAc-Et,O Ace-Et,O Ace-Et,O Ace Ace Ace Ace Ace-Et,O Ace-Et, 0 Ace
Formulab C,,H,,NO~HCl C,, HI,FNO. HCl C,,H,,ClNO~HCl C,,H,,BrNO~HCl C,,H,,F,NO~HCl C, H,, F,NO*HCl C, ,HI ,F, NO. HCl C,,H,, NO.HC1 C,,H,, NO; HC1 C,,H,, N0;HCl C,,H,,NO,~HCl C,,H,,NO,~HCl C,,H,,NO,~HBr C,,H,,NO,~HBr C,,H,, NO,.HCl C,,H,,N,O,~HCl C,,H,,NO,.HCl C,,H,,FNO;HCl C,,H,,ClNO; HCl C,,H,,BrNO;HCl C,,H,,F,NO,*HCl C, ,H,, F, NO,. HCl C, HI,F, NO; HCl C,,H,, NO,.HCl C,,H,, NO,. HCl C,,H,, NO,*HCl C,,H,,NO,*HCl C, ,H2,NO,. HCl C,,H,,NO;HCl C,,H, ,N, 0,. HC1
,
absorption. The methyl ketal of 2 was prepared (2 refluxed in CH30H-HC1 for 8 h) in order to prevent tautomeric interconversion. Comparison of the NMR spectral characteristics of 2 and its methyl ketal derivative in solvents of different polarity (CDC13and pyridine-&) and at variable temperature (25-100 "C) revealed no alterations of the signal patterns for the C(4)-methyleneprotons (AB q) or for the C(1)-methylene protons (AB pattern). These data suggest a similar conformational and configurational structure for the methyl ketal of 2 and, further, a homogeneity for the structure of the hemiketal2. The lack of a tautomeric interconversion in a series of 2-arylmorpholine hemiketals has also been reported for those derivatives containing a tertiary N.6 Similar spectral characteristics were noted for the substituted phenyl derivatives of 1 and 2, suggesting similar conformational and configurational preferences for all members of the series. Pharmacology. Earlier studies of the CNS effect of 1 and 2 indicated that these compounds were depressants of activity in mice in the 10-40-mg/kg range.3 Hence, the substituted aryl derivatives of 1 and 2 were also evaluated for effects on the spontaneous locomotor activity in mice as a quantitative measure of CNS effects. Selected members of each series (1 and 2) were screened for anticonvulsant activity in protecting mice from maximal electroshock or metrazole seizures. The neurotoxicity of these same compounds was also evaluated with a rotorod performance test. The effects of 1, 2, 8, 11, 23, and 26 on food consumption in rats were also evaluated in order to determine if these analogues possessed anorexic activity. These particular derivatives of 1 and 2 were chosen on the basis of their structural relationships with phendime-
462 Journal of Medicinal Chemistry, 1978, Vol. 21, No. 5
Rankin, Riley, Murphy
Table 11. Effects of 3-Aryl- and 3-Hydroxy-3-aryloctahydropyrido[ 2,1-c][ 1,410xazines on Spontaneous Locomotor Activity in Mice
X=OH X=H Redn of act. at Redn in act. at Compd R 40 mg/kg ipa Compd R 40 mg/kg ipa 1 H +++ 2 ++ 22 7 4-F -t+ +++ 23 8 4-C1 +++ t++ 24 9 4-Br +++ ++ 25 10 2-CF3 +++ ++ 26 11 3-CF3 +++ +++ 12 4-CF3 ++ + ++ 27 28 13 4-CH, + +++ 14 3-OCH3 i 29 ++ 15 4-OCH3 +i+ 30 + 31 ++ ++ 16 3,4-(OCH,), ++ 32 17 3,4,5-(OCH,), 18 3-OH 33 19 4-OH 34 +++ 20 3-OAc 21 4-NHAc ++t Chlorpromazine hydrochloride +++b (5 mg/kd Sodium phenobarbital C ( 2 0 mg/kg) a + + + , reduction in activity not significantly different from chlorpromazine hydrochloride ( 5 mg/kg, p < 0.05); + t , reduction in activity significantly different from saline (ip, p < 0.05); t, reduction in activity but within the standard deviaProduced Produced a 83.6% reduction in spontaneous locomotor activity. tion of saline; -, n o reduction in activity. approximately 215% increase in spontaneous locomotor activity.
trazine, chlorphentermine, and fenfluramine. Results and Discussion The effects of the substituted phenyl analogues of 1 and 2 on spontaneous locomotor activity in mice are summarized in Table 11. At the highest dose level tested (40 mg/ kg), those compounds exhibiting measurable activity were found to be depressants. As might be expected, relative lipophilicity appears to play a role in determining the potency of depression of activity as evidenced by the greater activity of the analogues of 1 as compared to 2 and by the enhancement of depressant activity produced by incorporation of a lipophilic moiety (halophenyl, trifluoromethylphenyl). This generalization, however, fails to explain the strong activity depressant effects of the 4’-OH (19) and 4’-NHAc (21) analogues. It also appears that positional isomerism in the phenyl group is an important structural effect, as evidenced by the differences in activity of the 3’- and 4’-OMe (14 and 15) analogues. These differences are minimal in the same analogues of the hemiketal 2 series. A number of the compounds studied exhibited a dose-dependent, biphasic response with regard to spontaneous locomotor activity. The 4’-trifluoromethylphenyl analogue 27 produced significant stimulation of activity a t 20 mg/kg and significant depression a t 40 mg/kg. However, the 4’-CF3 analogue of 1 failed to elicit a biphasic response. The 3’-hydroxyphenyl analogue 18 was a depressant of activity a t the 20 mg/kg level but not a t the 40 mg/kg dose. A similar profile of central activity is seen with certain hallucinogenic agents.12 The anticonvulsant activities of those analogues of 1 and 2 evaluated appear to be restricted to protecting test animals from major convulsions (maximal electroshock) (Table 111). In many cases, these analogues were found to potentiate metrazole-induced seizures. The most active compounds of this study were the trifluoromethylphenyl derivatives. The 3’-trifluoromethylphenyl analogues (11 and 26) exhibited the most favorable ED5” values which
were comparable to the EDs0 of phenobarbital sodium. Perhaps more important was the finding that the protective indices (TD50/ED,o)for 11, 26, and phenobarbital sodium, 3.11, 2.77, and 3.88,16respectively, are similar. The 4’-fluorophenyl analogue 7 was also of interest in that it was found to be a potent convulsant in untreated animals at doses of 100 mg/kg. Evaluation of potential anorexic actions for certain analogues of 1 and 2 indicated that compounds 1, 8, 11, and 26 produced a significant decrease ( p < 0.05) in food intake in rats a t the dose levels tested. Only the phenyl (1) and 4’-chlorophenyl (23) hemiketal derivatives were inactive in this assay. The doses (mmol/kg) of 1, 11, and 26 required to reduce food intake by 50% were 108.4,55.9, and 29.6, respectively, while for fenfluramine hydrochloride this dose was determined to be 16.8 mmol/kg. It is also pertinent t o note that these concentrations of 11 and 26 do not produce significant depression of activity in mice, whereas the anorexic activity of fenfluramine hydrochloride is associated with a depressant effect.13 The 4’-chlorophenyl analogue 8 was lethal in the test animals a t doses of 20 and 40 mg/kg but decreased food intake by 40% a t a 34.7 mmol/kg dose level. In summary, it appears that the most promising compounds prepared in this study are the 3’-trifluoromethylphenyl analogues of 1 and 2. These compounds are effective anticonvulsants and appetite suppressants at dose levels that do not appreciably depress motor activity in mice. Future studies will be directed toward optimizing potential therapeutic actions of appropriate analogues of 1 and 2 via systematic structure-activity relationship analyses. Experimental Section General. Melting points were determined on a Thomas-Hoover melting point apparatus and are uncorrected. Spectral data were obtained for all compounds and were utilized in confirming structure. The IR spectra were determined using a Beckman
Substituted Octahydropyrido[2,1-c][l,4]oxazines
Journal of Medicinal Chemistry, 1978, Vol. 21, No. 5 463
Table 111. Anticonvulsant Activities of Substituted Aryl Derivatives of 1 and 2 Lowest dose demonstrating act., mg/kg ip Antimax shock
Antimetrazole
22 10
100 100 ia 100 30-100
iab ia ia ia 300
25
30-100
ia
11
30-100
ia
26
30-100
ia
12
30-100
300
27
100
ia
13 28 29 30 31 32 Phenobarbital sodium
ia 100 100 ia ia ia
ia ia ia ia 100 300
Compd 1 2 7
Antimax shock,a ED,, mmol/kg
Ro torod toxici ty,a TD,, mmol/kg
258.9 (214.3-298.0) 186.0 (166.5-203.7 ) 84.6 (69.7-109.0) 78.7 (60.4-91.2) 290.8 (227.0-346.0) 202.1 (171.8-262.0)
715.7 (480.3-823.6) 546.6 (484.1-599.3) 262.6 (222.2-287.9) 217.9 (192.4-236.8) 407.8 (288.1-490.5) 468.OC (431.6-500.8 )
85.7d (59.0-1 00.3)
332.8d (29 5.8-41 7.0)
ia indicates the lack of anticonvulsant activity in doses of The figures in parentheses are 95% confidence intervals. 300 mg/kg o r less. This value is an LD, value for 28 due to the lethality produced by this compound. See ref 15. IR-33, except for dilution studies which were performed using a Perkin-Elmer Model 257 spectrometer using matched sodium chloride cells of widths 1.091 mm for 0.2 M solutions and 25 mm for 0.002 M solutions. The NMR spectra were taken using a Jeolco C-6OHL spectrometer. Mass spectra were obtained using a Du Pont 21-492 mass spectrometer a t 70 eV and using the electron-impact technique. Microanalyses are within *0.4% of theory and were performed by Chemalytics, Inc., Tempe, Ariz., and Galbraith Laboratories, Inc., Knoxville, Tenn. All compounds were tested as their HCl salts except for 18 and 19 which were tested as the HBr salts. Pharmacological Methods. Compounds subjected to the spontaneous locomotor screen were administered (ip) in three dosage levels (10, 20, and 40 mg/kg) in groups of eight male mice (20-30 g) per dose using distilled water as the vehicle. Test solutions were prepared to deliver the appropriate dose in 1 mL per 100 g of body weight. A group of eight male mice served as the control group and received 0.9% saline, 1 mL per 100 g of body weight. A 10-min stabilization period was employed immediately after administration of the test compound. The animals were then housed singly in photobeam cages containing six beams arranged in a circular fashion. Following a 1-min period during which the animals were not scored, the number of beam interruptions was determined at 10-min intervals for 30 min for both saline controls and treated groups. Significance of the data was determined through an analysis of variance, followed by application of the Duncan multiple range test.I4 Significant differences in effects on activity from saline ( p < 0.05) and insignificant differences in effects on activity from chlorpromazine hydrochloride ( p < 0.05) were evaluated for each compound at each dose level tested. Evaluation of anticonvulsant activity was performed in the anticonvulsant screening project by the Antiepileptic Drug Development Program (NIH). All compounds were tested in male Carworth Farms No. 1 mice and administered ip in three dose levels as the HC1 salts in 0.9% saline. Initial testing involved determination of the lowest of the three dose levels required to protect the mice from maximal electroshock and/or subcutaneous metrazole seizures (Table 111). Those compounds exhibiting anticonvulsant activity a t